Skip to content
AI-grafen
Sign in
☰ Menu

The goal F AI engineering

AI safety in practice

Red teaming, jailbreaks, adversarial examples, preference learning and reward hacking — what can be defended against and what cannot.

Knowledge nodes
65
From zero
about 57 h
Labs
6
See what you already know — no account

The diagnostic removes what you already know, so your path is usually much shorter.

What you can do afterwards

Labs along the way

You write the code. Tests you cannot see decide whether it holds up.

Lab: a neural network in pure NumPy — forward, backprop, gradient checkDa sandbox · about 75 minLab: dot product, norm and cosine similarityDin the browser · about 40 minLab: gradient descent from scratchDin the browser · about 50 minLab: matrix multiplication and one layer of a neural networkDin the browser · about 45 minLab: scaled dot-product attention with a causal maskDa sandbox · about 60 minLab: semantic search with embeddingsDa sandbox · about 45 min

The whole path

Everything the goal builds on, grouped by level and in the order it builds on itself. Show on the map

AExplorer2 knowledge nodes
  1. Patterns and categories
  2. Sequences and precise instructions
BInvestigator6 knowledge nodes
  1. Data in everyday life
  2. Algorithmic thinking
  3. Rule-based systems and machine learning
  4. Source criticism and responsibility in AI use
  5. Training data, features and labels
  6. Classification: how a model sorts information
CBuilder12 knowledge nodes
  1. Functions and coordinate systems
  2. Programming logic — variables, conditions, loops
  3. Python — the basics
  4. Python — lists, loops and dictionaries
  5. Python — strings and text processing
  6. Python — files, CSV and JSON
  7. Statistics — mean, median and spread
  8. Probability — the basics
  9. Linear regression: fitting a straight line to data
  10. Neural networks — the intuition
  11. Prompting — steering a language model
  12. Responsible use of AI
DAI developer22 knowledge nodes
  1. Derivatives and optimisation
  2. APIs and HTTP
  3. Regular expressions
  4. Licences and open data
  5. Loss functions
  6. Tokenisation
  7. Gradient descent
  8. The context window, system prompts and few-shot
  9. Vectors
  10. Matrices and matrix multiplication
  11. Linear regression with several features
  12. Neural networks — the forward pass with matrices
  13. Backpropagation
  14. Embeddings — words as vectors
  15. Attention
  16. Overfitting and generalisation
  17. Partial derivatives and the gradient
  18. PyTorch — tensors and autograd
  19. Retrieval — finding the right text
  20. Training, validation and test
  21. Train a neural network in PyTorch
  22. Transformers — the architecture
EUniversity10 knowledge nodes
  1. Reinforcement learning — the basics
  2. Convolutional networks (CNNs)
  3. Language models — training and generation
  4. Model evaluation
  5. Fine-tuning language models
  6. RAG — retrieval-augmented generation
  7. Prompt injection
  8. Tool use
  9. Agents — plan, act, observe
  10. Web scraping — the technique and the rules
FAI engineering11 knowledge nodes
  1. Adversarial examples
  2. Dataset design for fine-tuning
  3. Evals for language models and agents
  4. RLHF and preference learning
  5. Reward hacking
  6. DPO and direct preference optimisation
  7. Specification problems and proxy objectives
  8. AI safety and red teaming
  9. Jailbreaks and guard rails
  10. Training data for language models: filtering and dedup
  11. Data poisoning and backdoors
GFrontier Lab2 knowledge nodes
  1. Alignment — the problems and the methods
  2. Constitutional AI and rule-based alignment